151–160 of 824 posts
Earlier quoted context omitted.
> A programmer can read available but not oss licensed code and learn from it Actually, we were forbidden to look at open source code at Microsoft (circa 2009) because it might influence our coding and violate licenses.
Do the TypeScript team code with their eyes closed?
If github or google indexes source code using a neural net to help you find it, given a query, is that also illegal? If you think of copilot as something that helps you find code you’re looking for, is it all that different, and if so, why? In this case, wouldn’t the users of copilot be the ones responsible for any copyrighted code they may have accessed using copilot?
//below output code is MIT licensed (source: github/repo/blah)
And yes, the "users" are responsible, but it's possible that copilot could be implicated in a case depending on how it's access is licensed.
Stable diffusion has this same problem btw, but in visual arts "fair use" is even murkier.
For code, if you could use the code and respect the license, why wouldn't you? Copilot takes away that opportunity and replaces it with "trust us".
I'm kinda sceptical that this goes anywhere given that basically they say that whatever copilot outputs is your responsibility to vet that it doesn't break any copyright (obviously that goes against the promise of it and the PR but that's the small print that gets them out of trouble).
Saying "it's your responsibility to not breach licenses or violate copyright" doesn't absolve your service from breaching licenses and violating copyright itself.
Earlier quoted context omitted.
Saying "it's your responsibility to not breach licenses or violate copyright" doesn't absolve your service from breaching licenses and violating copyright itself.
Yet we all use web browsers that copy copyrighted text from buffer to buffer all the time. This doesn't even include all of the copying that ISPs perform. It might be fair to say that the read performed in training has the same character since no human is involved. The real copyright violation would be using a derived work.
when someone uploads their copyrighted text to a web page they are distributing it to whoever visits that page. the browser is just the medium.
As a non-lawyer, I am very suspicious of the claim that "Plaintiffs and the Class have suffered monetary damages as a result of Defendants’ conduct." Flagrant disregard for copyright? Sure, maybe. The output of the model is subject to copyright? Who knows! But the copyright holders being damaged in some what? Seems doubtful. The best argument I could think of would be "GitHub would have had to pay us for this, and th…
The common practice in copyright cases is to calculate damages based on the theoretical cost that the infringer would have paid if they have bought the rights in the first place. This method was used during the piratebay case to calculate damages caused by the sites founders. They did not actually calculate damages in terms of lost movie tickets or estimates vs actually sales number of sold game copies. When it came…
I am not against this lawsuit but I'm against the implications of this because it can lead to disastrous laws. A programmer can read available but not oss licensed code and learn from it. Thats fair use. If a machine does it, is it wrong ? What is the line between copying and machine learning ? Where does overfitting come in ? Today they're filing a lawsuit against copilot. Tomorrow it will be against stable diffusio…
If a human programmer reads some else's copyrighted code, OSS or otherwise, memorizes it and later reproduces it verbatim or nearly so, that is copyright infringement. If it wasn't, copyright would be meaningless.
The argument, so far as I understand it, is that Copilot is essentially a compressed copy of some or all of the repositories it was trained on. The idea that Copilot is "learning from" and transforming its training corpus seems, to me, like a fiction that has been created to excuse the copyright infringement. I guess we will have to see how it plays out in court.
As a non-lawyer it seems to me that stable diffusion is also on pretty shaky ground.
APIs are not copyrightable (in the US), so Wine is safe (in the US).
On page 18, they show Copilot produces the following code: >function isEven(n) { > return n % 2 === 0; >} They then say, "Copilot’s Output, like Codex’s, is derived from existing code. Namely, sample code that appears in the online book Mastering JS, written by Valeri Karpov." Surely everyone reading this has written that code verbatim at some point in their lives. How can they assert that this code is derived specif…
> Surely everyone reading this has written that code verbatim at some point in their lives Ironically their Twitter account uses a screenshot from a TV series as profile picture. I wonder how legal that is, even if meant as a joke. https://twitter.com/saverlawfirm Edit: It's been changed 2 minutes after I wrote this comment
Or is your comment itself the joke?
The value of copyleft licenses, for me, was that we were fighting back against the notion of copyright. That you couldn't sell me a product that I wasn't allowed to modify and share my modifications back with others. The right to modify and redistribute transitively though the software license gave a "virality" to software freedom.
If training a NN against a GPL licensed code "launders" away the copyleft license, isn't that a good thing for software freedom? If you can launder away a copyleft license, why couldn't you launder away a proprietary license? If training a NN is fair use, couldn't we bring proprietary software into the commons using this?
It seems like the end goal of copyleft was to fight back against copyright, not to have copyleft. Tools like copilot seem to be an exceptionally powerful tool (perhaps more powerful than the GPL) for liberating software.
What am I missing?
As a non-lawyer, I am very suspicious of the claim that "Plaintiffs and the Class have suffered monetary damages as a result of Defendants’ conduct." Flagrant disregard for copyright? Sure, maybe. The output of the model is subject to copyright? Who knows! But the copyright holders being damaged in some what? Seems doubtful. The best argument I could think of would be "GitHub would have had to pay us for this, and th…
The common practice in copyright cases is to calculate damages based on the theoretical cost that the infringer would have paid if they have bought the rights in the first place. This method was used during the piratebay case to calculate damages caused by the sites founders. They did not actually calculate damages in terms of lost movie tickets or estimates vs actually sales number of sold game copies. When it came…
The average salary of a programmer in which country?
So much programming is outsourced these days, and in some places programmers are very cheap.
It feels weird saying this but, for once, I hope the big evil corporation gets to keep selling their big bad product. I find the pattern matching and repetitive code generation really helpful. And the library autocomplete on steroids, too. Meh. Tricky subject.
That's where the line is for it to be suspect IMO.